Software development cost estimation – a field study
Cost estimation is an important practice in management accounting to support decision-making. In the IT-industry, however, software cost estimation (SCE) is a challenging field, because almost all of the costs in software engineering consist of labor effort and the understanding of cost drivers is still limited. Different estimation methods exist and are being put to use while none has prevailed to be the ‘best’. There is little guidance in the literature for companies which do not yet systematically estimate their software costs, and practical examples do not exist in the literature. This interventionist research endeavor in a software company in the health care sector aims to narrow this gap of knowledge by providing insights on how an estimation method can be identified, introduced and evaluated. For this purpose, the estimation context, the selection process of a SCE method, the implementation, and finally the evaluation of the introduced analogy approach are illustrated. During the course of this research project, a database of 358 past projects was set up and software cost drivers were identified and implemented within a similarity metric for purpose of estimating new development projects. Finally, the method was tested in practice. As a result, it seems that the selected estimation approach performed neither better nor worse than the subjective assessment of the experts in the company. However, the analogy method is able to systematically assess the reliability of its estimation which the specialists within the organization were not able to provide. While the estimation and its risk assessment seem to be promising, more field data needs to be collected in the future before coming to a final conclusion.